1. Field of the Invention
The present invention relates to a focus adjusting device, an image pickup apparatus, and a focus adjustment method. More particularly, the present invention relates to autofocusing used in, for example, an electronic still camera or a video.
2. Description of the Related Art
In an image pickup apparatus (typically, a digital camera), an autofocusing (AF) method making use of an object detection function of detecting, for example, the moving body or the face of a human being is well known. In the AF method, first, a main object is detected from a picked up image, to control focusing so that the detected main object is brought into focus.
In the AF method, when the main object is successfully detected, a focus detection area is set at the position of the detected main object, to control the focusing so the main object is brought into focus. However, when the main object cannot be detected, an area where the main object might exist, such as the center of a screen, is assumed, to set the focus detection area.
When the AF method is used to continue focusing the main object at all times at the focus detection area which is set as in, for example, a continuance AF operation or an AF operation during movie recording, the main object may or may not be successfully detected. In such a case, the position of the focus detection area is frequently changed between a main-object area and the central area of the screen. Here, when the main object does not exist at the central area of the screen, as shown in FIGS. 8A and 8B, the focus position moves between the main object and a background object, thereby resulting in flickering of the screen, which is troublesome. To overcome this problem, for example, it is necessary to set the focus detection area at the position of a previously detected main object, even if the main object cannot be successfully detected.
As a related detecting unit that detects a main object and that tries to solve to the aforementioned problem, a camera that is provided with a line-of-sight detecting function (in which the size of an area selected by line-of-sight detection is changed in accordance with the reliability of an output of the line-of-sight detection) is proposed (refer to Japanese Patent Laid-Open No. 6-308373). According to the document, when the reliability of the output of the line-of-sight detection is low, the size of a focus detection area is made large, to set a main object within the focus detection area.
In addition, a main object detecting unit that makes a focus detection area large when the main object cannot be detected is proposed in Japanese Patent Laid-Open No. 2004-138970.
Further, as an autofocusing controlling device using a moving-object detecting unit, a device that calculates the difference between AF evaluation values for image frames having different times for respective AF areas that have been previously divided and set is proposed. By calculating this difference, it is possible to extract an AF area including a moving object, so that the AF area is focused as an area to be focused. (Refer to Japanese Patent Laid-Open No. 2000-188713.)
However, Japanese Patent Laid-Open No. 6-308373 and Japanese Patent Laid-Open No. 2004-138970 discuss methods that only increase the focus detection area when the main object cannot be detected (that is, when the reliability of the detection of the main object is low). These methods do not allow the main object to be reliably set in the focus detection area unless, for example, the focus detection area is spread over the entire screen. In addition, when the focus detection area is merely increased, the proportion of the background in the focus detection area is increased. As a result, the background may be brought into focus.
In the case where a moving-object detecting unit, such as that discussed in Japanese Patent Laid-Open No. 2000-188713, is used, when the main object is a moving body, it is possible to always detect the main object to set the focus detection area. However, since a plurality of focus detection areas must be set over the entire screen, it takes time to perform calculate differences by an amount corresponding to the number of focus detection areas. Therefore, time is wasted in correspondence with the number of unnecessary focus detection areas. In addition, the main object cannot be detected when it does not move.